Description of problem: My sound card is not recognized for several kernels. the kernels I have on my pc are: 6.5.13 6.6.22 6.6.28 It's an comuter I don't use much and only when traveling, which is why I haven't done much updating. In particular, it was used during the ISO MGA9 test phases, when everything was functional. My hardware: processor Intel(R) Atom(TM) CPU Z3775 @ 1.46GHz bridge Atom Processor Z36xxx/Z37xxx Series SoC Transaction Register display Atom Processor Z36xxx/Z37xxx Series Graphics & Display bus Atom Processor Z36xxx/Z37xxx, Celeron N2000 Series USB xHCI usb1 bus xHCI Host Controller input ASUS Base Station(T100) bus T100TA usb2 bus xHCI Host Controller bus T100TA storage T100TA HDD DOCKING generic Atom Processor Z36xxx/Z37xxx Series Trusted Execution Engine bridge Atom Processor Z36xxx/Z37xxx Series Power Control Unit
Output of inxi -A
(In reply to katnatek from comment #1) > Output of > > inxi -A Audio: Message: No device data found. API: ALSA v: k6.6.28-desktop-1.mga9 status: kernel-api Server-1: PulseAudio v: 16.1 status: active (root, process)
In mcc -> hardware -> browse and configure hardware -> sound card can you see the device ? What was the last kernel the audio works?
(In reply to katnatek from comment #3) > In mcc -> hardware -> browse and configure hardware -> sound card can you > see the device ? > > What was the last kernel the audio works? I can't remember last kernel works. It works on MGA8 and during Alpha and Beta MGA9 tests. I can't see the device, I see device unknown: Atom Processor Z36xxx/Z37xxx Series Trusted Execution Engine Class media: Encryption-controller Module: mei_txe
Does it by chance work better using pipewire? https://wiki.mageia.org/en/Pipewire If not you could try our linus kernel (less patched by us)
CC: (none) => fri
I think we need the system journal as soon as possible after booting: # journalctl -b --no-hostname > afile.txt $ xz afile.txt This will produce afile.txt.xz; please attach that.
CC: (none) => lewyssmith
Created attachment 14529 [details] log journalctl
Thank you for that. Please say what sound card it is; and whether it is PCI or USB. I should have asked also for the O/P of: $ lspci $ lsusb (In reply to Guillaume Royer from comment #0) > It's a computer I don't use much and only when traveling, which is why I > haven't done much updating. > In particular, it was used during the ISO MGA9 test phases, when everything > was functional. So a definite regression.
lspci 00:00.0 Host bridge: Intel Corporation Atom Processor Z36xxx/Z37xxx Series SoC Transaction Register (rev 0d) 00:02.0 VGA compatible controller: Intel Corporation Atom Processor Z36xxx/Z37xxx Series Graphics & Display (rev 0d) 00:14.0 USB controller: Intel Corporation Atom Processor Z36xxx/Z37xxx, Celeron N2000 Series USB xHCI (rev 0d) 00:1a.0 Encryption controller: Intel Corporation Atom Processor Z36xxx/Z37xxx Series Trusted Execution Engine (rev 0d) 00:1f.0 ISA bridge: Intel Corporation Atom Processor Z36xxx/Z37xxx Series Power Control Unit (rev 0d) lsusb: Bus 002 Device 003: ID 0b05:17f9 ASUSTek Computer, Inc. T100TA HDD DOCKING Bus 002 Device 002: ID 0b05:17f6 ASUSTek Computer, Inc. T100TA Bus 002 Device 001: ID 1d6b:0003 Linux Foundation 3.0 root hub Bus 001 Device 003: ID 0b05:17f7 ASUSTek Computer, Inc. T100TA Bus 001 Device 002: ID 0b05:17e0 ASUSTek Computer, Inc. ASUS Base Station(T100) Bus 001 Device 001: ID 1d6b:0002 Linux Foundation 2.0 root hub
Thank you for those outputs. No sign of the missing sound card. > Please say what sound card it is; > and whether it is PCI or USB. Are you able to boot the computer from a Live USB, and see the sound card? If so, please provide the outputs of: 1. Either 'lsusb' or 'lspci' as appropriate. 2. $ inxi -Axx You understand we need to see what *should* be recognised & shown; what is not being seen.
I just remembered that I had a similar problem a few years ago, and it was TMB who solved the problem. I was missing the alsa-sof-firmware package. I'm reposting the bug report from that time for reference. I don't understand why this package was not installed during the installation of my system. Sorry for this bug report, I'm making you reinvent the wheel. https://bugs.mageia.org/show_bug.cgi?id=29705 The sound has now been restored.
Great the problem is found. So what caused the firmware to not be installed, is the question...
Re comment 11 - well done Guillaume. It will help re comment 11, 12 to know how you installed the system which had the problem: the missing package 'alsa-sof-firmware'. CC'ing Martin who knows about ISO contents.
Source RPM: Kernel 6.6.28 => Kernel 6.6.28, alsa-sof-firmware-2.2.5-1.mga9.nonfree.src.rpmSee Also: (none) => https://bugs.mageia.org/show_bug.cgi?id=29705CC: (none) => mageia
The alsa-sof-firmware package is on the classical installer ISOs. But with no sound device showing up in the output from lspci and lsusb, there may be nothing to trigger its installation. We'd need to see the installer logs (/root/drakx/report.bug.xz) to check that.
(In reply to Lewis Smith from comment #13) > Re comment 11 - well done Guillaume. > It will help re comment 11, 12 to know how you installed the system which > had the problem: the missing package 'alsa-sof-firmware'. > CC'ing Martin who knows about ISO contents. The installation of Mageia 9 was done from the beta. My computer had been used for testing the first ISOs. Then it stayed on the stable branch.
(In reply to Martin Whitaker from comment #14) > The alsa-sof-firmware package is on the classical installer ISOs. But with > no sound device showing up in the output from lspci and lsusb, there may be > nothing to trigger its installation. We'd need to see the installer logs > (/root/drakx/report.bug.xz) to check that. Thank you. So if the sound card had been seen during installation, the missing pkg would have been installed. It looks possible that the problem stems from an atypical installation, and would not happen on a virgin instal or upgrade. > We'd need to see the installer logs (/root/drakx/report.bug.xz) to check that. @Guillaume Please look at this file, and if its date looks sensible re the installation, please attach it to this bug.
Created attachment 14539 [details] Report bug
The file is dated February 2023, which is the correct installation date for my computer.
From the log, the driver for the sound device hasn't been loaded. Testing on my ASUS X205TA, which has the same Intel Bay Trail chipset, shows this is also true for the final Mageia 9 ISO. It looks like the installer image doesn't include the necessary kernel driver modules (snd-sof-acpi-intel-byt and dependencies). In addition, the entry in rpmsrate to include the firmware is 5 DRIVER"^snd_(hda_intel|sof_pci)" alsa-sof-firmware which would need updating to 5 DRIVER"^snd_(hda_intel|sof_pci|sof_acpi)" alsa-sof-firmware
Looking closer, the installer initrd images don't include any sound drivers. Sound card detection is done by matching the PCI or USB IDs. The difficulty is that on the Bay Trail platform, the sound device is neither PCI nor USB, and can only be detected by inspecting the ACPI tables. From reading the Linux kernel documentation, this is also true for the Haswell, Broadwell, and Cherry Trail platforms. Either we have to modify the installer to inspect the ACPI information, or we have to unconditionally install alsa-sof-firmware, or we have to leave this as an errata item.
Thank you Martin, and Guillaume for reporting. So installer/release media issue. Adding for errata 9 as the problem is in the current release. And target Mageia 10, hoping for fix in either way in mga10. - if not this need an errata entry every release. Suggestion on how to formulate the errata entry? I have not read up on this, so please improve my quick suggestion: " The installer fail to detect some sound hardware, and thus support for it is does not get automatically installed. This is true for Haswell, Broadwell, Bay Trail and Cherry Trail platforms. WORKAROUND: install the package alsa-sof-firmware and reboot. "
Keywords: (none) => FOR_ERRATA9Target Milestone: --- => Mageia 10Component: RPM Packages => Installer
Thaank you all for identifying & documenting an obscure but real problem. Luckily the fix is simple & clean. (In reply to Martin Whitaker from comment #20) > Looking closer, the installer initrd images don't include any sound drivers. > we have to unconditionally install alsa-sof-firmware, or we have to leave > this as an errata item. Can we do "unconditionally install alsa-sof-firmware"? And does the following also need doing? Or is this suggestion superseded by the that above? (In reply to Martin Whitaker from comment #19) > It looks like the installer image doesn't include the necessary kernel > driver modules (snd-sof-acpi-intel-byt and dependencies). In addition, the > entry in rpmsrate to include the firmware is > 5 DRIVER"^snd_(hda_intel|sof_pci)" alsa-sof-firmware > which would need updating to > 5 DRIVER"^snd_(hda_intel|sof_pci|sof_acpi)" alsa-sof-firmware
The solution is to install the alsa-sof-firmware package and reboot.(In reply to Morgan Leijström from comment #21) > Thank you Martin, and Guillaume for reporting. > > So installer/release media issue. > > Adding for errata 9 as the problem is in the current release. > > And target Mageia 10, hoping for fix in either way in mga10. > - if not this need an errata entry every release. > > Suggestion on how to formulate the errata entry? > I have not read up on this, so please improve my quick suggestion: > > " > The installer fail to detect some sound hardware, and thus support for it is > does not get automatically installed. This is true for Haswell, Broadwell, > Bay Trail and Cherry Trail platforms. WORKAROUND: install the package > alsa-sof-firmware and reboot. > " As for the errata, I'd put it like this: Some sound hardware is not detected by the installer and is not installed automatically. This is true for the platforms: Haswell, Broadwell, Bay Trail Cherry Trail Notably present on Intel Corporation Atom Processor Z36xxx/Z37xxx Series.
Thanks. Entered: https://wiki.mageia.org/en/Mageia_9_Errata#Sound
Assignee: bugsquad => isobuildSummary: No sound card recognize => No sound card recognizedKeywords: FOR_ERRATA9 => IN_ERRATA9
commit d3cdcb217498dff78dddb66547dad6c7b3aa4c9b Author: Martin Whitaker <mageia@...> Date: Sat May 25 10:43:19 2024 +0100 Detect legacy (non-PCI) Intel SST audio DSP devices (mga#33145) On the Broadwell. Bay Trail, Cherry Trail, and Braswell platforms the CPU has direct access to the SST sound device, so the device doesn't appear on the PCI bus. So the only way to detect these devices is to inspect the ACPI device list. As there are only a few devices involved, include the device table here rather than extending ldetect-lst. --- Commit Link: https://gitweb.mageia.org/software/drakx/commit/?id=d3cdcb217498dff78dddb66547dad6c7b3aa4c9b
This should have been assigned to mageiatools, not isobuild, but I've fixed it anyway. I tested it with a local rebuild of the mga9 installer. It will be available for testing in cauldron with the next release of drakxtools/drakx-installer-stage2 and meta-task.
Assignee: isobuild => mageiaSource RPM: Kernel 6.6.28, alsa-sof-firmware-2.2.5-1.mga9.nonfree.src.rpm => drakxtools-18.65-1.mga9, drakx-installer-stage2-18.66-2.mga9, meta-task-9-2.mga9Status comment: (none) => Fixed in git + svn